Participant attrition in a multi-regional population-based study in Africa: insights from the AWI-Gen study

Abstract

Attrition in longitudinal studies, defined as the loss of participants over time, may threaten the validity of study results. This study examined patterns, magnitude, and risk factors for attrition in the AWI-Gen cohort. The study included 12,032 participants from six African sites, with a median baseline age of 51 years (IQR: 45–57 years). Exploratory data analysis was conducted to examine attrition patterns and magnitude, and fixed-effects logistic regression models were used to identify significant factors associated with different causes of attrition (all-cause, non-death vs death, preventable vs non-preventable) while accounting for site-specific variation. Sensitivity analyses, including Bonferroni adjustments, imputation, and site-specific analyses, were used to assess model robustness. Model goodness-of-fit was evaluated, and adjusted odds ratios with 95% confidence intervals were reported, with a significance threshold of p < 0.05. After five years, AWI-Gen experienced 40% attrition. Causes included being unreachable by cell phone (42.6%), refusal (26.0%), death (13.9%), relocation (10.5%), and being recorded as too ill to participate (3.9%), with variations across sites. Male sex and lower socioeconomic status were consistently associated with increased odds across attrition outcomes, while older age, tobacco use, hypertension, chronic kidney disease, and cardiovascular disease were additionally associated with death-related and non-preventable attrition. This study underscores the importance of reporting attrition causes for consistency in longitudinal research. Male participants, individuals of lower socioeconomic status, older adults, tobacco users, and those living with chronic conditions such as hypertension, chronic kidney disease, and cardiovascular disease were at increased risk of attrition. Tailored retention strategies, including enhanced participant tracing and communication, targeted engagement of men and socioeconomically disadvantaged groups, regular contact with older participants, and integration of follow-up activities with chronic disease management services, may improve participant retention and help preserve the validity of longitudinal research.